In our previous research we developed highly conformal radiation treatment for head/neck and brain tumors using sculpted dose distributions and demonstrated the ability to escalate tumor dose with minimal complications (brain), and maintain high control rates while decreasing a major complication (xerostomia) for head/neck cancers. Our long-term goal is to increase tumor control and reduce complication rates by individualizing the timing and intensity of therapy based on innovative imaging. We propose to intensify treatment for glioblastoma by escalating the daily fraction doses while also utilizing effective chemotherapy (temozolomide). We will test the hypothesis that diffusion MRI (dMRI), which may predict tumor response early after therapy is started, will be a useful basis for re-optimization during the course of treatment, allowing the escalation of dose to non-responding parts of the tumor. We will also make use of our preliminary MRI results which suggest that radiation can open the bloodtumor- barrier (BTB) to chemotherapeutic and radiosensitizing agents (like gemitabine). We will test whether concurrent conformal radiation and gemcitabine (administered after the BTB is opened) will be safer and improve response rates for grade 3 gliomas, and whether we can improve the outcome by escalating the doses to the parts of the tumor demonstrating lack of increased permeability of the BTB early after the start of therapy. In head and neck cancer, we will test strategies that promise to decrease treatment-related late dysphagia and aspiration. We will test if reducing dose to the swallowing structures can reduce the severity of late dysphagia, and test the utility of innovative imaging (FLT-PET, and dMRI) to predict (early after the start of therapy) which tumors are likely to fail treatment, as well as directing individualized dose intensification using re-optimization plans. We expect that the Project will show improving non-complicated tumor control rates in brain and head and neck cancer by customizing treatment intensity following analysis of possible early predictors of response.